A language-independent LESK based approach to Word Sense Disambiguation

This paper describes a languageindependent LESK based approach to Word Sense Disambiguation (WSD), involving also Vector Space Models applied to the Distributional Semantics Hypotesis.In particular this approach tries to solve some issues that come up with less-resourced languages.The approach also addresses the inadequacy of the Most Frequent Sense (MFS) heuristics to fit specific domain corpora.